Promote an implicit property to an explicit axis.

Given an existing_axis with a property implicit_axis, create a new axis from the unique non-empty values of the property. Scan the other vector properties on existing_axis; for each one whose value is uniquely determined by the implicit value, migrate it to the new axis.

Usage

reconstruct_axis(
  daf,
  existing_axis,
  implicit_axis,
  rename_axis = NULL,
  empty_implicit = NULL,
  implicit_properties = NULL,
  skipped_properties = NULL
)

Arguments

daf

A DafWriter.

existing_axis

Axis that holds the implicit property.

implicit_axis

Property name on existing_axis; becomes the new axis's name (unless rename_axis).

rename_axis

Optional name for the new axis.

empty_implicit

If non-NULL, values equal to this are treated as empty (equivalent to the empty string).

implicit_properties

Optional character vector: only these properties are considered for migration.

skipped_properties

Optional character vector: properties to exclude from migration (even if consistent).

Value

Named list of "value for empty-implicit entries" per migrated property.

Details

Returns a named list: for each migrated property, the (consistent) value associated with existing_axis entries whose implicit value is empty — or NULL if no such entries exist. These values can be used to reconstruct the original property via the ?? X query modifier.

Requires that rename_axis (or the default, implicit_axis name) does not already exist in daf. Merging into a pre-existing axis is not supported.

Examples

d <- memory_daf(name = "d")
add_axis(d, "cell", c("c1", "c2", "c3"))
set_vector(d, "cell", "donor", c("dA", "dB", "dA"))
set_vector(d, "cell", "donor_age", c(30L, 40L, 30L))
reconstruct_axis(d, existing_axis = "cell", implicit_axis = "donor")
#> list()
get_vector(d, "donor", "donor_age")
#> dA dB 
#> 30 40